proflevelci: Produce cumulative sum credibility intervals

In ralmond/CPTtools: Tools for Creating Conditional Probability Tables

Produce cumulative sum credibility intervals

Description

Produces credibility intervals for hanging barplots. Assumes that each column represents a sum of proportions and produces corresponding intervals for the cumulative sums. Values hanging below and above the reference line are treated separately, and returned values below the reference are negative.

Usage

proflevelci(data, profindex, limits=list(lower=.025,upper=.975),a=.5, b=.5)

Arguments

data	A matrix of data values where each column refers to a bar in barplot. The values should be scaled so that the column sum is the number of individuals in that group.
profindex	The level in the chart which corresponds to the reference (proficiency) line. This should be a positive integer less than nrow(data).
limits	The upper and lower credibility limits.
a	Value for the shape1 parameter of the beta prior.
b	Value for the shape2 parameter of the beta prior.

Details

For a stacked bar plot, the natural comparisons involve not category probabilities but the sum of the category probabilities up through the current bar. For hanging bar plots, this should go in both directions. So for example, if the categories are “Below Basic”, “Basic”, “Proficient”, and “Advanced”, and the zero line is to be put between “Basic” and “Proficient”, then we need credibility intervals for Pr(“Basic” or “Below Basic”), Pr(“Basic”), Pr(“Proficient”), Pr(“Proficient” or “Advanced”).

The proflevelci function splits the states up into those above the line and below the line using profindex. It then generates credibility intervals using betaci for the cumulative sums in each group. The primary purpose is to create confidence intervals for stacked bar charts (see compareBars2).

Value

A list of data sets of the same length as the limits argument. Each data set has the same shape as the data argument and represents a quantile of the data associated with the value in limits. With the default limits of lower and upper, the result is a list of two elements

lower	Gives the lower bounds of the confidence interval.
upper	Gives the upper bounds of the confidence interval.

Author(s)

Russell Almond

See Also

betaci, compareBars2

Examples

margins <- data.frame (
 Trouble=c(Novice=19,Semester1=24,Semester2=28,Semseter3=20,Semester4=9),
 NDK=c(Novice=1,Semester1=9,Semester2=35,Semseter3=41,Semester4=14),
 Model=c(Novice=19,Semester1=28,Semester2=31,Semseter3=18,Semester4=4)
)

proflevelci(margins,3,limits=c(lower=.025,upper=.975))

ralmond/CPTtools documentation built on Dec. 27, 2024, 7:15 a.m.